Repository: https://github.com/Fua6655/rur_ros
Platform: SoftBank Robotics Pepper
Framework: ROS (Robot Operating System)
📺 Demo video: https://www.youtube.com/watch?v=QCsrS-MGEd8
📄 IEEE publication: https://ieeexplore.ieee.org/document/8870908
This project implements a ROS-based multi-robot system for both
Human–Robot Interaction (HRI) and Robot–Robot Interaction (RRI),
developed and deployed in a live theatrical performance based on
R.U.R. (Rossum’s Universal Robots).
Robot behavior is designed using a formal behavior model based on Finite State Machines (FSM) and Behavior Tree principles, enabling robust, predictable, and safety-aware execution in a real-world environment.
The system was executed on physical Pepper robots, interacting with human actors, other robots, and the audience in real time.
A recording from the live performance demonstrating:
- human–robot interaction
- robot–robot coordination
- real-time execution on stage
👉 https://www.youtube.com/watch?v=QCsrS-MGEd8
✔️ Live theater performance
✔️ Human actors and audience present
✔️ Multiple Pepper robots
✔️ Real-time execution (no simulation)
✔️ Safety-critical interaction
This deployment required:
- reliable behavior transitions
- synchronization between robots
- predictable responses to humans
- tolerance to timing and environmental uncertainty
The system supports structured interaction between robots and humans:
- Interaction with human actors on stage
- Behavior transitions triggered by human cues
- Expressive motion, posture and timing
- Safety-aware behavior execution in close proximity to humans
Robots were required to:
- maintain safe distances
- behave predictably for human performers
- integrate interaction timing with the theatrical script
The project also implements robot–robot interaction, enabling:
- Coordination between multiple Pepper robots
- Synchronization of behaviors and actions
- Inter-robot communication via ROS topics
- Shared state awareness across robots
This transforms the system from isolated robot scripts into a distributed multi-agent robotic system.
Robot behavior is implemented using a structured behavior model inspired by:
- Finite State Machines (FSM) for clear state transitions
- Behavior Tree concepts for modular, hierarchical behavior design
This approach enables:
- deterministic and explainable robot behavior
- safe transitions between interaction states
- reuse of behavior modules
- robustness in live, unstructured environments
The FSM/Behavior Tree model proved essential for:
- handling interaction timing
- managing concurrent human and robot interactions
- avoiding unsafe or undefined robot states
The system is built on a modular ROS architecture:
- ROS nodes for behavior execution
- Topic-based inter-robot communication
- Configuration-driven state transitions
- Real-time message handling
This architecture supports:
- scalability to additional robots
- reuse in non-theatrical HRI scenarios
- adaptation to other human-centered robotic applications
rur_ros/ ├── launch/ # ROS launch files ├── nodes/ # Interaction & behavior nodes (FSM / BT logic) ├── scripts/ # Helper scripts ├── config/ # YAML behavior and state configs ├── docs/ # Design notes and planning └── README.md
✔️ ROS-based HRI & RRI system
✔️ Multi-robot coordination
✔️ FSM / Behavior Tree behavior modeling
✔️ Real-time execution
✔️ Safety-aware interaction logic
✔️ Live deployment with human presence
This project demonstrates advanced robotics competencies:
- Human–Robot Interaction (HRI)
- Robot–Robot Interaction (multi-agent systems)
- Formal behavior modeling (FSM / Behavior Trees)
- ROS ecosystem proficiency
- Real-world deployment under uncertainty
It shows the ability to design robust autonomous behavior, not just scripted motion.
This work directly aligns with EU priorities in:
- human-centric robotics
- collaborative and social robots
- safe autonomous systems
- AI & robotics integration
It is particularly relevant for EU Cascade Funding calls (e.g. MAGICIAN, euROBIN) and short-term robotics R&D contracts.
Luka Kicinbaci — Robotics / ROS Developer
GitHub: https://github.com/Fua6655
Developed and deployed a ROS-based multi-robot system implementing both human–robot and robot–robot interaction, using a behavior design based on FSM and Behavior Tree principles, executed in a live theatrical performance with human actors and audience.